#!/usr/bin/env python
# coding: utf-8

# In[3]:


import numpy, scipy, matplotlib.pyplot as plt, sklearn, stanford_mir
get_ipython().run_line_magic('matplotlib', 'inline')


# [← Back to Index](index.html)

# # Cross Validation

# <a href="https://en.wikipedia.org/wiki/Cross-validation_(statistics)">K-fold cross validation</a> is a method for evaluating the correctness of a classifier.
# 
# For example, with 10-fold cross validation:
# 
# 1. Divide the data set into 10 random partitions.
# 2. Choose one of the partitions as the test set. Train on the other nine partitions.
# 3. Repeat for the partitions.

# Load some features from ten kick drums and ten snare drums:

# In[4]:


training_features, training_labels, scaler = stanford_mir.get_features()


# In[13]:


print training_labels


# Plot their features:

# In[2]:


plt.scatter(training_features[:,0], training_features[:,1])


# Initialize the classifier:

# In[28]:


model = sklearn.neighbors.KNeighborsClassifier(n_neighbors=3)
model = sklearn.linear_model.LogisticRegression()


# Perform 5-fold cross validation:

# In[29]:


acc  = sklearn.cross_validation.cross_val_score(model, random_features, training_labels, cv=5)


# In[30]:


print acc.mean()


# In[ ]:


k


# [&larr; Back to Index](index.html)